Body Surface Context: A New Robust Feature for Action Recognition From Depth Videos.
Yan SongJinhui TangFan LiuShuicheng YanPublished in: IEEE Trans. Circuits Syst. Video Technol. (2014)
Keyphrases
- action recognition
- view invariant
- spatio temporal interest points
- human actions
- motion features
- body parts
- action classification
- recognition of human actions
- video dataset
- depth sensors
- recognizing human actions
- mid level features
- activity recognition
- bag of words
- static images
- recognizing actions
- human detection
- space time interest points
- ucf sports
- human activities
- action recognition in videos
- action detection
- depth cameras
- computer vision
- single view
- bag of features
- robust tracking
- global features
- video frames
- multi view
- human body
- human object interactions
- d objects
- low level descriptors
- three dimensional
- feature descriptors
- motion history images
- motion capture data
- human motion
- surface reconstruction
- depth information
- visual words
- range data
- image sequences
- dynamic scenes
- machine learning